| Interface | Description |
|---|---|
| DirectConfirmationMeasure |
This interface is implemented by all confirmation measures which are using
the probabilities directly.
|
| LogBasedCalculation |
This interface defines the epsilon used by log based calculations for
avoiding the logarithm of 0.
|
| Class | Description |
|---|---|
| AbstractUndefinedResultHandlingConfirmationMeasure |
An abstract class of all confirmation measures which have to handle undefined
results.
|
| CondProbConfirmationMeasure |
This confirmation measure calculates the conditional probability of W' given
W*. result = P(W',W*)/P(W*)
|
| DifferenceBasedConfirmationMeasure |
This confirmation measure calculates the difference between the conditional
probability of W' given W* abd the marginal probability of W'. result =
P(W'|W*)-P(W')
|
| FitelsonConfirmationMeasure |
This confirmation measure calculates the difference of the conditional
probability of W' given W* and the conditional probability of W' given ¬W*.
|
| JaccardConfirmationMeasure |
This confirmation measure calculates the Jaccard similarity between W' and
W*. result = P(W',W*)/P(W' v W*)
|
| LikelihoodConfirmationMeasure |
This confirmation measure calculates the Likelihood similarity between W' and
W*. result = P(W'|W*)/(P(W'|¬W*) + e)
The e is defined by
LogBasedCalculation.EPSILON. |
| LogCondProbConfirmationMeasure |
This confirmation measure calculates the logarithm of the conditional
probability of W' given W*. result = log(P(W',W*)/P(W*))
|
| LogJaccardConfirmationMeasure |
This confirmation measure calculates the logarithm of the Jaccard similarity
between W' and W*. result = log((P(W',W*) + e)/P(W' v W*))
The e is defined by
LogBasedCalculation.EPSILON. |
| LogLikelihoodConfirmationMeasure |
This confirmation measure calculates the Likelihood similarity between W' and
W*. result = (P(W'|W*) + e)/(P(W'|¬W*) + e)
The e is defined by
LogBasedCalculation.EPSILON. |
| LogRatioConfirmationMeasure |
This confirmation measure calculates the logarithm of the ratio between the
joint probability of W' and W* and the product of the two marginal
probabilities. result = log((P(W',W*) + e)/(P(W')*P(W*)))
Another name for it is pointwise mutual information (PMI).
|
| NormalizedLogRatioConfirmationMeasure |
This confirmation measure calculates the logarithm of the ratio between the
joint probability of W' and W* and the product of the two marginal
probabilities.
|
| OlssonsConfirmationMeasure |
This confirmation measure calculates Olssons coherence.
|
| RatioConfirmationMeasure |
This confirmation measure calculates the ratio between the joint probability
of W' and W* and the product of the two marginal probabilities. result =
P(W',W*)/(P(W')*P(W*))
|
| ShogenjisConfirmationMeasure |
This confirmation measure calculates Shogenjis coherence.
|
Copyright © 2014–2016. All rights reserved.